> We can even add more real information and still have less data to encode

The reason they're so verbose comes from a few facts.

First of all, it's an existing logistics labelling standard, they've just replaced brackets with forward slashes and put a domain name on the front. So https://example.com/01/09521207311511/21/1234ABDE1235 is just a QR code version of those huge barcodes like (01)09521207311511(21)1234ABDE1235 you see on cases of products in the supermarket.

Second of all, the standard doesn't limit itself to a single date, so they can't identify dates with a simple ,D prefix. It's a kitchen sink standard [1] with 16 different types of date (production date, due date, packaging date, sell by date, best before date, expiration date, release date, first freeze date, harvest date, production date and time...) and just as many options for sizes and weights - so the identifier can be up to 4 digits. /11/ or /8008/

The third thing to know is QR codes pack different alphabets at different densities. Numbers at 3.5 bits per character, upper case letters and some symbols at 5.5 bits per character, ASCII at 8 bits per character. So the 13 characters of of "Apple Fuji Sm" uses about as much space in a QR code as a 29-digit number like "12345678901234567890123456789"

Fourth, you've replaced the 14-digit GTIN with an 11-digit UPC and replace the 13-digit serial number with a 5-digit lot code :)

IMHO the standard isn't going to take over the world, and anyone who says "Barcodes are about to go extinct" doesn't know what they're talking about. But it's not the information density, it's other reasons.

[1] https://ref.gs1.org/ai/